The health consequences of diet don’t largely depend on whether a person eats a high or low level of carbohydrates or a diet high or low in fat. What’s much more important is where the carbs and fats come from, according to an analysis that related diet and mortality rates in more than 37,000 American adults.
“Unhealthy low carbohydrate diet [LCD] and low-fat diet [LFD] scores were associated with higher total mortality, whereas healthy LCD and LFD scores were associated with lower total mortality,” Zhilei Shan, MD, and associates wrote in an article (JAMA Intern Med. 2020 Jan 21; doi: 10.1001/jamainternmed.2019.6980). The findings “suggest that the association of LCDs and LFDs with mortality may depend on the quality of food sources of macronutrients,” said the researchers, based at the Harvard T.H. Chan School of Public Health in Boston.
The analysis included follow-up of almost 300,000 person-years. It showed that, for every 20-percentile increase in a person’s unhealthy LCD score, their relative rate of total mortality increased by a statistically significant 7%; and for every 20-percentile rise in unhealthy LFD score, the relative, total mortality rate rose by a statistically significant 6%, after adjustment for several demographic and clinical measures and family and personal histories of diabetes, cancer, and heart disease. In contrast, for each 20-percentile increase in healthy LCD score relative, total mortality fell by 9%, and similar increases in healthy LFD score linked with an 11% relative drop in total mortality, also statistically significant associations in these confounder-adjusted analyses.
The findings “extend the previous evidence” for these links, and the data suggest that “the health benefits of an LCD or LFD may depend not only on the types of protein and fat or carbohydrate but also on the quality of carbohydrate or fat remaining in the diet,” the researchers wrote. They cited the documented health problems caused by eating significant amounts of low-quality carbohydrates such as refined grains and added sugars, which provide limited nutrition and introduce a high glycemic load, and can produce high levels of postprandial glucose and insulin, inflammation, insulin resistance, and dyslipidemia.
The foods people ate that produced healthy diet scores and linked with better survival were diets high in plant protein and unsaturated fat, and low in carbohydrates from refined grains, added sugar, starchy vegetables, and similar sources as part of a low carbohydrate diet. The foods that formed a healthy LFD included whole grains, whole fruit, legumes, and nonstarchy vegetables, along with higher intake of plant protein and low levels of saturated fat.
The study used data from 24-hour diet-recall surveys completed by 37,233 American adults collected by the National Health and Nutrition Examination Survey (NHANES) during 1999-2014, and linked the diet scores calculated for these people with U.S. national death records collected by the National Death Index through the end of 2015. The people included averaged about 50 years of age at the time of their dietary interview, and 53% were women. During 297,768 person-years of follow-up, 4,866 total deaths occurred, including 849 from heart disease and 1,068 from cancer. The analyses found no statistically significant links between overall LCD or LFD scores and mortality; the significant links only existed when the researchers further classified the diet scores into healthy and unhealthy subtypes.
The results also showed statistically significant links or strong trends between high or low levels of healthy or unhealthy LCD and LFD scores and cancer deaths. A 20-percentile increase in unhealthy LCD score linked with an 11% relative increase in cancer deaths, while a 20-percentile increase in the healthy LCD score linked with a 10% decrease in cancer deaths. A 20-percentile increase in the healthy LFD score linked with a 15% relative decrease in cancer mortality.
The study received no commercial fundings, and the authors had no commercial disclosures.
SOURCE: Shan Z et al. JAMA Intern Med. 2020 Jan 21; doi: 10.1001/jamainternmed.2019.6980.